Mollie and Emil will have their daily workplace at the School of pharmacy/Harvard School of Public Health and Institute of Mathematics, respectively. Below, they have presented themselves.
Post doc Mollie Wood – started July 1st in PharmaTox
Photo: private
I am very excited to be joining the PharmaSafe group full time!
I recently completed my PhD in Clinical and Population Health at the University of Massachusetts Medical School.
My thesis assessed the risk of neurodevelopmental problems in children following prenatal exposure to triptans, using multiple causal inference approaches, including marginal structural models, propensity score matching and calibration, mediation analysis, and probabilistic bias analysis. I found that triptan exposure during pregnancy, particularly in the first trimester, was associated with externalizing behaviors but not with psychomotor function or internalizing behaviors; various methods used to adjust for unmeasured confounding by indication suggest that underlying migraine severity does not fully explain these results.
I am interested in applying rigorous methodological approaches to assessing medication safety during pregnancy; in particular, my research interests include causal inference methods for assessing time-varying exposures and mediators, as well as methods for dealing with bias arising from misclassified exposure and from unmeasured confounding.
In my free time, I enjoy scuba diving, swimming, and playing the mandolin.
PhD student Emil Aas Stoltenberg – starting 1st August 2015
Photo: private
My academic background is in the social sciences and statistics.
In the spring of 2013 I submitted my master thesis at the Department of Political Science. This spring (2015) I submitted my thesis in statistics at the Department of Mathematics.
In the first thesis I attempted to predict the outcome of Norwegian parliamentary election of 2013, using a model that combined economic variables, polling data as well previous elections. The thesis in statistics is devoted to the simultaneous estimation of multiple Poisson means. I develop a class of estimators that perform uniformly better (in terms of risk) than the maximum likelihood estimator when estimating an ensemble of Poisson means. The class of estimators is constructed so that the statistician can compromise between the aim of good ensemble properties (all the means) and good individual properties (when focusing on individual means).